One hundred and eighty-six previously untreated patients with clinical stage D2 prostate cancer have been followed according to the criteria of objective response of the National Prostatic Cancer Project (NPCP). All patients received combination therapy with the antiandrogen Flutamide and the LHRH agonist (D-Trp6, des-Gly-NH2(10)]LHRH ethylamide (or surgical castration, 10 patients) as first treatment. Forty-nine patients (26.3%) achieved a complete response as best response while 56 (30.1%) and 69 (37.1%) patients had partial and stable responses, respectively, and only 12 patients (6.5%) did not respond to treatment. The median times required to achieve stable, partial and complete responses were 155, 183 and 401 days, respectively. The best response achieved has a major influence on the probability of continuing response and survival. While the 50% probability of continuing response is more than 3 years for the complete responders, it is reduced to 630 and 517 days for partial and stable responders, respectively. While the non-responders have a median life expectancy of 10.0 months, this value is increased to 30.3 and 37.8 months for the stable and partial responders, respectively. The best probability of survival is for the complete responders with a 95.9% probability of survival at 3 years. There is no significant correlation between the time required to achieve a best response (phase 1) and the duration of the response before progression occurs (phase 2) or the time between progression and death (phase 3) for any of the categories of responses. A longer period of time required to achieve a complete response is associated with a longer survival. When analysis is made, in an attempt to predict response, of the baseline characteristics of the patients before treatment, a low number of bone metastases and better performance status are associated with a greater chance of achieving a complete response while partial, stable and progression responses cannot be predicted from the baseline characteristics. The present data show the importance of standardization of the objective criteria of response to treatment in advanced prostate cancer. Thus, the patients who achieve a complete response have a much more favorable prognosis while partial and stable categories of response have a closely similar prognosis which is inferior to the complete responders. Moreover, the present data indicate that the stable category of response has an important prognostic value which is almost superimposable and not statistically different from the partial response in terms of duration of response and survival.(ABSTRACT TRUNCATED AT 400 WORDS)